Interval ridge regression (iRR) as a fast and robust method for quantitative prediction and variable selection applied to edible oil adulteration
详细信息 查看全文
- 作者：Ozren Jović ; ozren_jovic@yahoo.com" class="auth_mail" title="E-mail the corresponding author ; ozren.jovic@chem.pmf.hr" class="auth_mail" title="E-mail the corresponding author ; Neven Smrečki ; Zora Popović
- 关键词：Ridge regression ; Partial-least squares regression ; Variable selection ; Hempseed oil ; Adulteration
- 刊名：Talanta
- 出版年：2016
- 出版时间：1 April 2016
- 年：2016
- 卷：150
- 期：Complete
- 页码：37-45
- 全文大小：1087 K
文摘
- •
constant λ iRR was successfully applied to FTIR, UV–vis, DSC and GC data.
- •
iRR significantly outperforms ridge regression and partial-least square regression.
- •
iRR is faster than iPLS and prevents possible under- and overfitting of iPLS.
- •
FTIR, UV–vis and DSC can determine adulteration of hempseed oil (RMSEP<2%).
- •
FTIR-iRR can attain very high accuracy and very low limit of detection.